Final report on data discrepancies in the CWT Tool study 1
Transcription
Final report on data discrepancies in the CWT Tool study 1
Report on data discrepancies in the CWT Tool – Study 1 Main authors: Dr Penny Newman, Amanda Dell, Ayo Adebamowo Contributors: Ian Craddock, David Traynier, Orla Thunder, Michelle Fisher, Lorna Dewar, Hether Buckle, Sandra Lyons, Dr Mary McStay, Dr Devy Basu, Dr Jennifer Collins, Denise Gale Executive Sponsor: Dr Sean MacDonnell This report was commissioned by Colchester Hospital University NHS Foundation Trust at the request of the multiagency incident management team (IMT). Dr Christine Macleod, on behalf of the Assurance Panel of the IMT, endorses the findings within this report. Signature: Date: 16/12 /2014 Dr Christine Macleod, Medical Director NHS England, Essex Area Signature: Date: 16/12 /2014 Dr Lucy Moore, Chief Executive Colchester Hospital University Foundation Trust. 1 15th December 2014 Our gratitude goes to the following: Colchester Hospital University NHS Foundation Trust Main authors: Dr Penny Newman (Review Programme Director, Director of Service Integration, GP and Consultant in Public Health), Amanda Dell (Project Management Consultant), Ayo Adebamowo (Lead Cancer Information Analyst) CHUFT Contributors and Review Team members: David Traynier (Audit support/Analysts), Lorna Dewar (Oncology Clinical Trials Manager), Orla Thunder (Clinical Trials Data Nurse), Michelle Fisher (Research Nurse), Ian Craddock (Site Matron, Senior Auditor), Heather Buckle, Sandra Lyons (Clinical Auditors), Dr Mary McStay, Dr Devy Basu, Dr Jennifer Collins (Lead Clinicians), Denise Gale (Cancer Programme Director), Valerie Northcroft-Brown (Project Support) Executive sponsor: Dr Sean MacDonnell (Medical Director) Business Informatics team and in particular Michele Figg (Head of Business Informatics) Health Records team and especially Barry Moult (Head of Information Governance & Health Records),Teresa Frost (Health Records Manager) and Phil Frances (Health Records Clerk) The Assurance Panel Dr Christine Macleod (NHS England Essex Area Team Medical Director and Chair of Retrospective Assurance Panel), Dr Shane Gordon (Clinical Chief Officer, NHS North East Essex Clinical Commissioning Group), Karen Hindle (Senior Associate (communications) interim hub manager, North, Midlands and East Communications), Dr Thomas Nutt (Chief Executive Officer, Healthwatch Essex), Paul Pharoah, (Professor of Cancer Epidemiology, Department of Public Health and Primary Care, University of Cambridge and Honorary Consultant in Public Health, Public Health England), Pól Toner (NHS England Essex Area Team Director of Nursing and Quality). Observer: Representatives of the Information Commissioner’s Office, Support: Christine Cooper (PA to NHS England Essex Area Team Medical Director) External reviewers The East of England Strategic Clinical Network, particularly Dr Rory Harvey (SCN Cancer Clinical Director) and Kate Patience (Rehabilitation & Quality Improvement Lead for East of England SCN Cancer Team), all the external review Consultants and Karen Harland (Cancer Services Manager, HOPE Clinical Unit, Hinchingbrooke Health care NHS Trust) The Royal Marsden NHS Foundation Trust and particularly Nicky Browne (Director of Performance & Strategy Implementation) 2 15th December 2014 Table of Contents Table of Contents .................................................................................................................................... 3 Executive Summary............................................................................................................................... 5 Abbreviations and Acronyms Used in this Report .................................................................................. 8 1. Introduction ................................................................................................................................... 9 1.1 The Retrospective Review Team and reporting arrangements .................................................. 10 1.2 The Assurance Panel ................................................................................................................... 10 2 National and local data on cancer waiting times ...................................................................... 11 2.1 Cancer Waiting Times Guidance ................................................................................................. 11 2.2 Comparative data........................................................................................................................ 12 2.3 Cancer mortality statistics and survival rates for Colchester Hospitals ...................................... 13 3 Aims and objectives..................................................................................................................... 14 4 Methods ........................................................................................................................................ 14 4.1 Audit process .............................................................................................................................. 14 4.2 Quantitative analysis................................................................................................................... 15 4.3 Analysis of cancer waiting time data .......................................................................................... 15 4.3.1 Analysis by cancer pathway (operational standard) ............................................................ 15 4.3.2 Analysis by data field ........................................................................................................... 15 4.3.3 Unverified data .................................................................................................................... 16 4.3.4 Discrepancies between the case notes and CWT tool ......................................................... 16 4.4 External review ........................................................................................................................... 16 4.5 Qualitative analysis ..................................................................................................................... 16 4.6 Queries and areas of concern ..................................................................................................... 17 4.7 Assurance panel review of potential data manipulation and/or areas of clinical concern ........ 17 5. Results .......................................................................................................................................... 18 5.1 Sample and patient characteristics ............................................................................................. 18 5.1.1 Source of Referral ................................................................................................................ 19 5.1.2 Shared Pathways .................................................................................................................. 19 5.2 Data accuracy .............................................................................................................................. 20 5.2.1 Inaccuracy by cancer pathway ............................................................................................. 20 5.2.2 Data inaccuracy by tumour site ........................................................................................... 21 5.2.3 Spread of data errors by tumour site................................................................................... 22 5.3 Causes of data errors .................................................................................................................. 23 5.4 Impact of data inaccuracy ........................................................................................................... 24 5.4.1 Time between recorded and actual dates ........................................................................... 24 3 15th December 2014 5.4.2 Significance of variance between CWT tool and data in the notes ..................................... 27 5.4.3 Clinical pathways and care ................................................................................................... 29 5.5 Correlation with Patient surveillance log .................................................................................... 29 6. Royal Marsden Hospital Review (Appendices 7 and 8) ........................................................... 31 7. Areas identified for improvement from qualitative analysis ................................................... 32 7.1 Poor record-keeping ................................................................................................................... 32 7.2 Decentralised record-keeping ..................................................................................................... 32 7.3 Misunderstanding of Cancer Waits guidance ............................................................................. 32 7.4 Misunderstanding of “Decision to Treat” ................................................................................... 32 7.5 Poor information sharing with other providers .......................................................................... 32 8. Discussion and conclusion .......................................................................................................... 33 9. Recommendations ....................................................................................................................... 34 9.1 ..................................................................................................................................................... 34 9.2 ..................................................................................................................................................... 34 9.3 ..................................................................................................................................................... 34 9.4 ..................................................................................................................................................... 34 9.5 ..................................................................................................................................................... 35 9.6 ..................................................................................................................................................... 35 9.7 ..................................................................................................................................................... 35 Appendix 1: Study Protocol ................................................................................................................ 36 Appendix 2: Data on Cancer Survival, Mortality and Referrals ....................................................... 39 Appendix 3: Methodology for Obtaining Sample Data .................................................................... 41 Appendix 4: Audit Tool and Fields ..................................................................................................... 43 Appendix 5: Glossary .......................................................................................................................... 47 Level One Codes:............................................................................................................................... 47 Level Two Codes: .............................................................................................................................. 47 Level Three ........................................................................................................................................ 48 Decision to Treat Time-only codes.................................................................................................... 48 Final codes ........................................................................................................................................ 49 Appendix 6: Variance (in days) for Data Items with Incorrect Dates ............................................. 50 Appendix 7: The Royal Marsden NHS Foundation Trust independent review of Colchester University Hospital NHS Foundation Trust Retrospective Review Team audit results ................ 51 Appendix 8: Marsden review of cases for Colchester Hospital NHS Foundation Trust as part of retrospective review – cases where there was disagreement with the Trust................................ 52 4 15th December 2014 Executive Summary Following concerns about cancer care at the Trust, in 2013 Colchester Hospital University NHS Foundation Trust (CHUFT) was inspected by the Care Quality Commission (CQC), who reported inconsistencies in data held on the Cancer Waiting Times system (CWT) and within 61 medical records. In 22 of these cases, the delay in treatment may have had an impact on patients’ health (see below). “During our examination of 61 patient medical records we noted there to be inconsistencies between information that was held on the cancer wait times system (CWT) and that which was contained within people’s medical records. In total 22 of the files showed discrepancies in a person’s pathways. We found that entry dates on the CWT system did not always correlate with the patients' medical records. We found that in 22 cases the treatment dates recorded on the system had been changed. Details and examples of the altered records are reflected in the section for 'records' in this report. The changes to the patient cancer pathway was identified through a review of the medical records which demonstrated that treatment had been provided on a date different to the one recorded on the CWT system. The changes in those 22 cases meant that people could have experienced a negative impact in the form of their treatment being delayed. The risk of delayed treatment could impact on their care and longer term health”. CQC Report into Cancer Services at Colchester Pages 12 and 7, Published 5th November 2013. The Retrospective Review was commissioned by NHS England and the Incident Management Team (IMT) as part of a suite of measures arising from the inspection and review of cases. The review aims to investigate, through a transparent audit process, the extent of data inaccuracies in Cancer Waiting Times in the Trust and their impact on clinical care. This report describes findings from Study 1 into the error rate of cancer pathway data and is one chapter of six in the Retrospective Review’s overall report. The Retrospective Review looked at two aspects of patient care at the Trust and any relationship between them: 1. The Cancer Waiting Times data pathway to identify data inaccuracy 2. The clinical pathway to identify any potential clinical harm. This study on data error rate looked more specifically at data issues. Its aim was to assess the historical accuracy of the data held on the CWT through case note review of a stratified sample of 250 records between April 1, 2010 and October 31, 2013 so that accurate Trust cancer waiting time performance can be assured. The audit was a six stage process including data sampling, template development, case note review, expert cancer wait pathway and clinical review, analysis and report writing and an external review of a random 10% of records. This was undertaken by a cancer data expert from the Royal Marsden NHS Foundation Trust to check auditors’ accuracy of assessment and validity of results and external Consultants identified by the Anglia Cancer Network to assess quality of care. Cases of “potential data manipulation” were reviewed against criteria by an Assurance Panel made up of all key 5 15th December 2014 stakeholders and chaired by NHS England Medical Director to identify any need for referral to the police. No data inaccuracy was found in 97 of the 250 cases (39%) reviewed and in 14 cases (6%)1 the dataset could not be entirely verified as the activity was at other Trusts. The audit indicates that, in 55% (139) of cases, across most specialties reviewed and for a number of reasons, there were data inaccuracies in the CWT database. For all patients on a cancer pathway, reviewers considered up to nine data fields, and there was an error rate of 10% in all completed data entries (142 of 1,453 cases). Although the actual error rate of 55% may not be representative of the whole CWT database, as the sample was taken from a selected cohort of specialties with known issues, the finding that over half of cases reviewed had at least one data inaccuracy indicates poor data recording and quality. As the upkeep e.g. legibility, filing, and maintenance of data in the case notes was often poor, and as there were multiple cancer databases, information was difficult to find. There were multiple causes of data error, which was categorised by IMT and are defined below: No errors The record is found to be accurate and complete in all fields. Data entry errors When CWT and other sources of information, including 40% (100 patient notes, match although the data has not been cases) updated e.g. on another pathway, appointment brought forward, incorrect treatment entry When an incorrect entry appears to have been made as a 11% (28 cases) result of a misunderstanding of the national CWT guidelines. Misinterpretation of the national guidance Operating process issue Data with other provider 39% (97cases) A substantive error in a process or procedure, as opposed to 3% (7 cases) the recording of that process or procedure. Data available on the CWT but could not be checked against 6% (14 cases) data in the notes which were held elsewhere. Other causes included poor information sharing, document handling and recording i.e. documents mislaid or incorrectly reported, and mis-recording of Decision to Treat (DTT) following discussion or Multi-Disciplinary Team (MDT) meeting. There were 105 instances or data points along the Cancer Waiting Times pathway (“episodes”), relating to 83 cases, where auditors recorded a difference (or variance) between the dates in the CWT and the notes. These relate to four fields: referral received date; date first seen, decision to treat, and treatment start date. 1 While 8% of the patients were treated in other Hospitals, it was possible to verify some records as information was available on RTT. Therefore only 6% of cases could not be verified due to shared care. 6 15th December 2014 Comparing the dates recorded in the notes and on the CWT tool: In most instances where an error was identified the difference between the date recorded on the CWT and the date in the notes was small. In 58% of episodes (61 of 105) the date recorded in the notes was within five days before or after the date recorded CWT Tool Dates recorded in the notes for all pathways range from between 77 days before to 52 days after dates recorded in the CWT tool. In the majority of episodes (65% or 68 of the 105 data points) the date recorded in the notes was earlier than the date recorded on the CWT Tool. For 10 patients, dates in the notes were between 10 and 52 days after dates recorded on the CWT tool, indicating that, for some patients, and especially as the pathways may be cumulative, there would have been unacceptable delays. This study shows pockets of poor data recording, particularly in Upper and Lower GI and Urology, especially around treatment dates. The size of the variance was larger in Lower GI and Urology, although numbers are small. On statistical analysis of the cases where there was a discrepancy in dates between the notes and CWT there is some evidence of systematic bias making waits seem shorter than in reality. This could have resulted from multiple causes as described above including erroneous interpretation of complex CWT guidance or even subconscious bias. The methods of this study did not allow an assessment whether this led to gain, for example, in recording of performance. However, in the four individual cases assessed by the Assurance Panel in which auditors found data irregularities, the Panel found no evidence of data manipulation. In these cases which were assessed to determine if potential data manipulation might have led to delays harmful to patients, there was no reason to refer any cases to the police. In the sample of 250 cases, there were five cases referred for Serious Incident investigations (SI). Two new referrals came from the retrospective review and three had already been referred previously via the helpline, complaints, or the SI process. These latter three SIs related to aspects of care other than those covered by the Error Rate audit. It should be noted that, during the period of the helpline, the threshold for SI investigations was lowered. One case was considered by the Assurance Panel a GP significant event (the equivalent to a Serious Incident but in primary care). After reviewing these SIs, the Trust put new policies in place and reinforced aspects of the Cancer Action Plan. As one significant event result from a GP referral on to a non-cancer pathway, and 5% of GP cancer referrals were routine, the Trust needs to work closely with primary care to improve early recognition of cancer and referral across the system. The Trust needs to work with other hospitals to improve data sharing as, in 14 of 21 cases of shared care, insufficient data were available to complete the audit. The picture is of a problem with poor data quality in cancer. Data from other Trusts have not been analysed to such a degree as a comparator. It is hoped that data recording will be addressed through the introduction of the Somerset Cancer Registry information system. In conclusion, although discrepancies in the recording of data were found in this study, there was no evidence of this causing patient harm. 7 15th December 2014 Abbreviations and Acronyms Used in this Report 2WW Two Week Wait IMAS Interim Management and Support CCG Clinical Commissioning Group IMT Incident Management Team CHUFT Colchester Hospital University Foundation Trust IT Information Technology MD Medical Director CI Confidence Interval MDT Multi-Disciplinary Team CNS Clinical Nurse Specialist MDTC Multi-Disciplinary Team Coordinator COSD Cancer Outcome and Service Dataset CQC Care Quality Commission MEHT Mid-Essex Health Trust NAO National Audit Office CUP Cancer of Unknown Primary NE North East CWT Cancer Waiting Times NEE North East Essex DNA Did Not Attend NHS National Health Service DTT Decision to Treat ONS Office for National Statistics ECAD Earliest Clinically Appropriate Date ECRIC Eastern Cancer Registration and OPD Outpatients’ Department PAS Patient Administration System Information Centre PH Public Health (England) FDT First Definitive Treatment RMH Royal Marsden Hospital FT Full Time RRT Retrospective Review Team GDP General Dental Practitioner RTT Referral to Treatment GI Gastro-Intestinal SCN Strategic Clinical Network GMP General Medical Practitioner SCR Somerset Cancer Registry GP General Practitioner TYA Teenagers and Young Adults SI A definition of a serious incident is: ‘An incident of such seriousness that it causes or threatens to cause serious harm to patients, staff, volunteers, members of the public, contractors or the Trust itself. There are specific criteria used to define SIs and this includes the risk of reputational harm to the NHS. A Never Event is a Serious Incident which should never happen if known preventive measures had been in place’(Series Incident Procedure Trust Policy No 63a, 2014) SI Cont. Please note where an SI is mentioned throughout all reports in the retrospective review this refers to an incident report form (datix) being completed and referral for serious incident investigation. The investigation process can take up to 45 days at which point a final report is submitted to North East Essex CCG for review. Some investigations are still on-going at time of writing this report. 8 15th December 2014 1. Introduction Following concerns about cancer care, in 2013 Colchester Hospital University NHS Foundation Trust (CHUFT) was inspected by the Care Quality Commission, which identified 30 patients who may have experienced harm resulting from potential data manipulation. As a result, the Trust reviewed all 30 cases and met with patients and their families. Where consent was given, cases were referred for medico-legal review. In addition, the Trust undertook a look-back exercise to address specific areas of concern identified in clinical pathways to ensure patients were safe. The Retrospective Review was commissioned by NHS England and the Incident Management Team (IMT) as part of a suite of measures including: A review of cancer pathways by the East of England Cancer Network An independent review of Trust Governance The implementation of a comprehensive Cancer Action Plan by the Trust including a new information system (Somerset Cancer Registry). The review aims to investigate, through a transparent audit process, the extent of data inaccuracies in cancer waiting times in the Trust and their impact on clinical care. It will recommend necessary remedial action to ensure accurate reporting and improved systems of care, and provide assurance to the public and stakeholders. More specifically the retrospective review will: 1. Using clinical and non-clinical audits, identify and report on the extent, at an individual and system level, of the following: - Evidence of data manipulation, - Causation factors, and where possible, whether these were intentional, - Whether patients have been exposed to any risk of harm, - Remedial actions required, including the on-going management and monitoring of cancer services. 2. To provide transparency and external validation through regular reporting to a Retrospective Review Assurance Panel and by publishing results. 3. To refer to the police any cases of potential manipulation of data. The Retrospective Review involves the study of six specific aspects of cancer care across all specialties and a review over 1,700 records: Study 1 To determine the error rate of the Cancer Waiting Tool (CWT) Study 2 Long waits - to identify long waits (over 91 days) in cancer. Study 3 To determine the prevalence of ‘delayed diagnoses’ in cancer pathways (defined as ‘those patients restarting a cancer pathway within 90 days from stopping an initial cancer pathway’). Upper GI - to identify patients whose pathway was changed without appropriate clinical input. Urology - to identify patients who have been lost to superficial bladder cancer surveillance. Surveillance – review of all calls to the helpline, complaints and significant events following the CQC investigation Study 4 Study 5 Study 6 250 case notes 290 case notes 364 case notes 120 case notes 15 case notes 684 patients 9 15th December 2014 All studies look, to varying degrees, at two aspects of patient care at the Trust and any relationship between them: 1. The Cancer Waiting Times data pathway to identify data inaccuracy 2. The clinical pathway to identify any potential clinical harm. This, Study 1 on error rate, looks in more detail at the data element. The protocol developed by IMT is in Appendix 1. 1.1 The Retrospective Review Team and reporting arrangements The Retrospective Review Team is composed of clinical auditors (two research nurses, a research manager, three general nurses or 4.42 Whole Time Equivalent (WTE)), an analyst (interim, Full Time, (FT), an audit facilitator (FT), a project manager (interim, FT) and a Programme Director (a GP and Consultant in Public Health, whose role was previously Director of Service Integration, 0.6 WTE). The team receive support on an ad hoc basis with CWT guidance from the Cancer Programme Director and clinical advice from three Consultants in Upper GI and oncology. The Programme Director reports to the Trust Medical Director (MD), Programme Steering Group and Turnaround Programme Board, and indirectly to the Executive Team. 1.2 The Assurance Panel Throughout the process, the Trust has reported to the Assurance Panel. This panel consists of all stakeholders (NHS England, North East Essex Clinical Commissioning Group (CCG), and representatives of the Information Commissioner, Healthwatch Essex, and Department of Public Health and Primary Care, Cambridge University). Its role is to oversee and validate the process and the results of the Retrospective Review, in order to provide assurance to the public, patients, staff and relevant bodies on data quality and clinical care in cancer at CHUFT. More specifically, it acts to: Review and advise on methodology to ensure it is transparent, robust and objective, Validate the accuracy of results, Gain assurance in the use of external consultants in the process for external validation, Comment on, inform, and sign off, the final report. 10 15th December 2014 2 National and local data on cancer waiting times 2.1 Cancer Waiting Times Guidance All work was undertaken are in reference to Cancer Waiting Times Guidance V8.0.2 This defines important terms, of which several are pertinent to this report. Decision to Treat (DTT): This is defined in the CWTs Guidance as “the date the patient agrees a treatment plan”. Two-Week Wait (2WW): This is defined as “urgent GP (General Medical Practitioner (GMP) or General Dental Practitioner (GDP) referral for suspected cancer to first outpatient attendance”. Patients referred on this pathway are required to be seen in the Trust within fourteen calendar days from the receipt of their referral. Patients on this pathway are concurrently on both the 62-Day pathway (from receipt of referral) and the 31 day 1st definitive treatment pathway (from Decision to Treat). 31-Day First Definitive Treatment (DTT): This is defined as “decision to treat to first definitive treatment”. Patients referred under this standard are required to commence first definitive treatment for a new cancer diagnosis within 31 days of the date of decision to treat being made. 62-Day Standard: This is defined as “urgent GP (GMP or GDP) referral for suspected cancer to first definitive treatment”. Patients referred with suspected cancer under the two-week wait standard (2WW) are required to commence first definitive treatment, if cancer is diagnosed, within 62 days from the date of receipt of referral for the suspected cancer, and within 31 days of the Decision to Treat. The 62-Day Screening: This is defined as “urgent referral from NHS Cancer Screening Programmes (breast, cervical and bowel for suspected cancer to first definitive treatment”. These patients are required to commence first definitive treatment, if cancer is diagnosed, within 62 days from the date of receipt of referral for the suspected cancer. 31-Day Subsequent Treatment: This is defined as “decision to treat/earliest clinically appropriate date (ECAD) to start of second or subsequent treatment(s) for all cancer patients including those diagnosed with a recurrence where the subsequent treatment is surgery, anti-systemic cancer treatment (drugs), or radiotherapy. These patients are required to commence treatment within 31 days of the date of decision to treat being made or earliest clinically appropriate date. Waiting Time Adjustment (First Seen): This records the number of days that patients should have been removed from their calculated waiting time for the two week wait period and potentially the 62 day period (if cancer is confirmed). Waiting Time Adjustment (Treatment): This data item is used to record the number of days that should be removed from the calculated waiting time between the date of decision to treat and the treatment start date i.e. the number of 2 http://www.nwlcn.nhs.uk/Downloads/Cancer%20Intelligence/Going%20Forward%20on%20Cancer%20Waits %20A%20Guide%20Version%208.0.pdf 11 15th December 2014 days that a clock can be paused for a 31 or 62 day period if a reasonable offer of treatment in admitted care has been declined. 2.2 Comparative data It is difficult to assess the significance of data error rate as comparative assessments of cancer information systems have not been undertaken. This means there is a lack of benchmark data against which to compare results. Nationally, there is a published error rate of 57% (incorrectly recorded or no evidence) in waiting times data (not cancer specific).3 The National Audit Office (NAO), NHS Waiting Times Elective Care in England was published in January 2014. Findings from the report are reproduced below. The NAO, NHS Waiting Times Elective Care in England Executive Summary Page 7 and 8 There are errors in the trusts’ recording of patients’ waiting time. We reviewed 650 orthopaedic patient waiting times across seven trusts. More than half of these were not supported by documented evidence or were incorrectly recorded. Although it was not a representative sample for the country as a whole, we established clear data risks that need to be managed. We found that: in 281 (43%) cases waiting times had been correctly recorded and were supported by documented evidence in 202 cases (34%), waiting times were not supported by enough evidence to say whether they had been correctly recorded; and in 167 cases (10%), there was evidence of at least one error, leading to under- and overrecording of waiting time. There was an overall under-recording of three weeks (mean) per patient, with a median of 11 days (paragraphs 2.14 to 2.18). Mis-recording of data was identified at The North West London, Barnet and Chase Farm and Colchester Trusts. The North West London Hospitals NHS Trust identified that it had failed to record properly the waiting times of 2,700 (60 per cent) of its elective (prearranged) inpatients, including 12 who had waited more than 52 weeks for treatment. Barnet and Chase Farm Hospitals NHS Trust identified that it had failed to monitor more than 2,000 patients on the waiting list, 651 of which had waited between 18 and 51 weeks for treatment (paragraphs 2.20 to 2.22). NAO Report Paras 2.0 to 2.22 The North West London Hospitals NHS Trust failed to record the waiting times of 2,700 (60 per cent) of its elective inpatients, including 12 patients who had waited more than 52 weeks for treatment. The errors were identified during an internal validation of inpatient waiting lists and reported to the trust’s board. In May 2013, Barnet and Chase Farm Hospitals NHS Trust identified a failure to monitor more than 2,000 patients on the waiting list, of which 651 had waited between 18 and 51 weeks. The backlog developed as a result of a failure in an IT reporting system. 3 http://www.nao.org.uk/report/nhs-waiting-times-elective-care-england/ 12 15th December 2014 At the time of the report to the board in May, there were no patients identified as waiting more than 52 weeks for treatment. However, further validation of the waiting list in July by trust staff identified 108 patients that had not been tracked and who had waited more than 52 weeks for treatment; these patients were subsequently reported to NHS England. In December 2013, Barnet and Chase Farm Hospitals NHS Trust had 181 patients who had waited more than 52 weeks for treatment. Responding to whistle-blowers, the Care Quality Commission reported that Colchester Hospital University NHS Foundation Trust had altered patient appointment and medical records on its cancer waiting times system. The Care Quality Commission found that in 22 cases the treatment dates recorded on the system had been changed. The police are now conducting an investigation at Colchester (paragraph 2.22). The report recognises that some of the challenges facing trusts when managing waiting lists are the perennial systemic issues of balancing financial and clinical capacity with the demand for services. 2.3 Cancer mortality statistics and survival rates for Colchester Hospitals The following statistics (Tables 1-5, appendix 2) are taken from the ONS, the National Cancer Registry, the Public Health England Data and Knowledge Gateway, the National Cancer Intelligence Network E Atlas: • • • • • • NE Essex CCG one-year survival index (%) is within normal range for all cancers combined for adults (aged 15-99 years) In 2011 NE Essex CCG one-year survival index was 67.7% for all cancers combined for adults (aged 15-99 years), slightly lower than all Essex CCGs combined (68.2%) and England as a whole (68.2%) although higher than three other CCGs in Essex NE Essex one and five-year survival rates for cancers within key specialties are all within the normal range NE Essex mortality 2008-2012 by cancer type are all within normal range except prostate, lung and stomach which are significantly lower than UK average Cancer incidence each year equates to around 0.5% of population for NE Essex higher than national rate of 0.37%, and has increased over last 13 years Total number of patients referred for 2WW is increasing although not translating into numbers diagnosed and treated. 13 15th December 2014 3 Aims and objectives The full protocol developed by IMT is in Appendix 1. Study Title: To determine the error rate of the CWT Tool Aim of the study: The aim of this study was to assess the historical accuracy of the data held on the Cancer Waiting Tool so that accurate Trust cancer waiting time performance can be assured. Primary Objective: To determine the historical accuracy of the data on the Cancer Waiting Tool. Secondary Objective: To estimate the impact of the CWT data accuracy on cancer waiting time performance reporting during the period of study Outcome Measures/Endpoints % data accuracy of CWT April 2010 – October 2013 Estimated impact on historic performance reporting 4 Methods The audit reviewed a stratified sample of 250 cancer records from between April 1, 2010 and October 31, 2013. The sampling process is detailed in Appendix 2 and includes a random sample specifically selected from four specialties with known issues; Breast, Lower Gastro-Intestinal (Lower GI), Upper Gastro-Intestinal (Upper GI), Urology and, to a lesser degree, other cancer pathways Inclusion Criteria 1. Randomised sample of 250 case notes (comprising 50 case notes from Breast, Lower GastroIntestinal (Lower GI), Upper Gastro-Intestinal (Upper GI), and Urology, and 50 from other cancer pathways), 2. Patients on either a 31-day or 62-day cancer pathway, 3. Patients on specified pathway between April 1, 2010 and October 31, 2013, 4. Patients who have had a first or second treatment. Exclusion Criteria 1. Where a patient was on multiple pathways, only one pathway was selected, 2. Patients who were not treated, 3. Patients who were discharged with no cancer diagnosis, 4. Patients who were not on a cancer pathway during the specified period, 5. Patients selected for previous audits. 4.1 Audit process The Retrospective Review process included six stages: Stage 1: Stage 2: Stage 3: Stage 4: Stage 5: Stage 6: Identifying the sample. Developing the audit tool and piloting Undertaking the case note review Clinical and cancer wait expert advice and/or review if necessary Analysis and report writing External review 14 15th December 2014 4.2 Quantitative analysis As the auditors reviewed each case, they validated pre-populated information recorded in the CWT Tool against data held in the Trust’s information systems in order to identify inaccuracies and variances. The information systems included the Patient Administrative System (PAS), case notes (general hospital notes and oncology notes), clinic letters, and pathology, radiology, and radiotherapy databases, as well as the national Open Exeter system. Additional information was sourced by contacting relevant clinical staff by phone or email. Where patients care was shared with other Trusts, CWT data was mostly unavailable. The data fields in the audit tool are given in Appendix 3. In addition to using categories established by IMT, each case note was reviewed using a set of predefined options or codes developed by the review team that allowed critical findings to be quantified and increasing levels of detail. A glossary of codes and definitions is given in Appendix 4. 4.3 Analysis of cancer waiting time data Analysis of the data recorded on the audit template comprised a descriptive analysis of all 250 cases and analysis of the congruence between the CWT database and notes by pathway type as a binary (correct/incorrect) or quantitative answer (variance in days) in up to nine data fields. 4.3.1 Analysis by cancer pathway (operational standard) The 250 cases audited included patients on different cancer pathways (Table 1). Patients can be on one or more pathway e.g. patients referred on a 62 day standard cancer pathway (via 2WW) will concurrently be on a 31 day 1st treatment pathway (once Decision to Treat has been confirmed). 2WW 31-Day 1st Treatment 31-Day Subsequent Treatment 62-Day Standard 62-Day Screening 36% (91 cases) 70% (175 cases) 30% (75 cases) 36% (91 cases) 10% (26 cases) Table 1: Distribution of sample of cases across CWT pathways (operational standards) 4.3.2 Analysis by data field The accuracy of data recorded in the Trust’s CWT tool was established in up to nine fields in each patient’s pathway. These were: Source of Referral, Date Referral Received, Date First Seen, Waiting Time Adjustment (First Seen), Date of Diagnosis, Decision to Treat Date, Treatment Start Date, Waiting Time Adjustment (Treatment), and Treatment Modality. The data items available for each case review were dependent on those required for submission according to CWT guidelines and pathway type i.e. whether a 2WW, 62-Day, Subsequent treatment: Of the nine fields, only four are mandatory for all four pathways and could therefore be analysed for all 250 patients. These fields include Date of Decision to Treat, Treatment Start Date, Waiting Time Adjustment (Treatment) and Treatment Modality. Data submission on Date of Diagnosis was not required for data returns until mid-2013 and therefore data was incomplete. The remaining four data items - Source of Referral, Referral Received Date, Date First Seen, and Waiting Time Adjustment (1st Seen) are mandatory only for patients on 62-Day Standard and 62-Day Screening pathways (117 patients). 15 15th December 2014 4.3.3 Unverified data Finally data analysis was dependent on checking data in the CWT against data in the notes. In the nine fields on average in 6% or 95 data points there was no other data to substantiate the CWT (Table 2). Of these, 25% (24 of 95) were due to missing data at CHUFT and 75% due to data held at other Trusts, equivalent to 1.5% and 4.5% of total data points respectively. Data Item No. of eligible No. of verified No. of data items % unable to data items data items unable to verify verify Source of Referral 117 109 8 7% Date Referral Received 117 108 9 8% First Seen Date 117 108 9 8% Waiting Time Adjustment 117 117 0 0% (First Seen) Date of Diagnosis 80 75 5 6% Decision to Treat Date 250 214 36 14% Treatment Start Date Waiting Time Adjustment (Treatment) Treatment Modality Total data items 250 250 234 249 16 1 6% 0% 250 1548 239 1453 11 95 4% 6% Table 2: Number of data items that could not be verified 4.3.4 Discrepancies between the case notes and CWT tool Analysis was undertaken on each of the 1,548 data points to determine the number and percentage of correct entries, incorrect entries, the size of discrepancy, and unverifiable entries (Table 2). For incorrect entries, a further level of analysis was done. In all nine fields, data error was recorded as a binary correct/incorrect answer while in five, auditors recorded the size of any discrepancy between dates in the notes and CWT database (referral received date, date first seen, decision to treat and treatment start date) and treatment modality. The data was analysed using Microsoft Excel™. 4.4 External review A random sample of 10% of audited records were assessed by a CWT expert from the Marsden assessed to check accuracy of CWT assessment and hence validity of results (reported on page 28) and by external Consultants from multiple specialties recruited by the Anglia Cancer Network to assess quality of care and appropriateness of subsequent action taken by the Trust as part of the review. 4.5 Qualitative analysis A thematic analysis was undertaken of free text comments for cases coded as for “non-clinical action” to identify areas that need to be addressed. Non-clinical action is defined as action taken to change administrative, technical, or clerical activity over and above the care of an individual patient. In addition, as they progressed, auditors listed broad themes and recommendations in an “Issues Log”. These issues form the basis for some of the Recommendations. 16 15th December 2014 4.6 Queries and areas of concern Where auditors had queries or concerns around data management and interpretation of the CWT guidelines, the Trust’s Cancer Programme Director was consulted to provide clarification of CWT guidance (v8.0). Where there were issues or concerns around patient care, either the lead clinicians working with the Retrospective Review Team were asked to assess the case to make recommendations for action or the case referred as a Serious Incident. 4.7 Assurance panel review of potential data manipulation and/or areas of clinical concern Following each case review, the auditors wrote a case summary. Together with the notes these were discussed at the Assurance Panel on 20th May, 2014 to review any cases of possible clinical harm and/or potential data manipulation. Cases were reviewed by the Assurance Panel against four questions to identify if any necessitated referral to the police. Potential data manipulation was defined as where there is no valid explanation why the CWT tool had been altered to meet national reporting standards, data did not reflect actual patient experience, there was variance between CWT and other data sources, there was no valid reason and/or other sources gave rise for concern e.g. complaints. 17 15th December 2014 5. Results The following summarises the key findings from the study, supplemented by qualitative analysis of the auditors’ comments from the audit tool and Issues Log. In some places, the text is supplemented by brief explanations of relevant Cancer Waiting Times guidance and codes from the Glossary (Appendix 4). The results are given in the following sections 1. 2. 3. 4. Sample and patient characteristics Data accuracy Causes of errors Impact of data inaccuracy 5.1 Sample and patient characteristics The 250 records audited comprised 50 case notes drawn at random from each of the following five tumour sites: Breast, Lower GI, Upper GI, Urology, and Others. The latter comprised the following numbers of case notes: Gynaecology (5), Head & Neck (3), Haematology (16), Lung (16), Cancer of Unknown Primary (CUP) (2), and Skin (8) (Figure 1). In the sample, 51% (127) of the patients were male and 49% (123) were female (Figure 2). The majority of patients were aged over 65 (67% or 168 patients), slightly higher than national cancer statistics where 63% are over aged 65 years (Figure 2).4 While the majority of patients in the sample were alive (72% or 180) at the time the audit was completed in April 2014, 28% (70 patients) were deceased. No. of Case Notes Audited Per Tumour Site 50 40 30 20 10 0 Breast Lower GI Upper GI Urology Others (Haema) Others (Lung) Others (Skin) Others (Gynae) Others (H&N) Others (CUP) Figure 1: Distribution of Case Notes Audited per Tumour Site 4 Figures from Cancer Research UK http://www.cancerresearchuk.org/cancerinfo/cancerstats/incidence/age/#Cancer 18 15th December 2014 Age Distribution 20% 18% 16% 14% 12% 10% 8% 6% 4% 2% 0% 25-29 30-34 35-39 40-44 45-49 50-54 55-59 60-64 65-69 70-74 75-79 80-84 85-89 90+ Figure 2: Age Distribution of Patients in Data Sample 5.1.1 Source of Referral The patients were referred to the Trust through different routes. 36% were referred from their GP via 2WW; 5% were routine referrals from GPs; 30% were referrals made by Consultants for subsequent treatment; 15% were referrals made by Consultants for investigations for cancer; 10% were referrals from a National Screening Programme, and 3% were referrals via emergency admission (Figure 3). Referral Source 45% 40% 35% 30% 25% 20% 15% 10% 5% 0% 104 75 37 26 7 GP Subsequent Treatment Consultant National Screening Programme Emergency Admission 1 Specialist Nurse Figure 3: Referral Source 5.1.2 Shared Pathways While the majority (92% or 229 patients) were treated at CHUFT, for 8% (21) of patients, the pathways originated from CHUFT but included treatment at Ipswich Hospital NHS Trust (7 cases) and Broomfield Hospital (14 cases). Of all patients, 70% (175) received first definitive treatment at CHUFT, while 30% (75) received second or subsequent treatment, perhaps because the Trust is a tertiary centre for radiotherapy. 19 15th December 2014 Patients Treated in Other Hospitals 35% 30% 7 6 25% 20% 15% 3 10% 3 2 5% 0% Lower GI Breast Others (Gynae) Others Upper GI Others (H&N) (Lung) Urology Others Others (Haema) (CUP) Others (Skin) Figure 4: Patients Treated at Other Hospitals by Speciality 5.2 Data accuracy Nine key data items were audited critical to capturing, tracking, measuring timelines of patients on cancer pathways, and assessing CWT performance (Appendix 3). Overall, data inaccuracies were found in at least one of the nine key data items in 139 of 250 cases reviewed (55%). No data inaccuracy was found in 97 of the 250 cases (39%) reviewed and in 6%5 data was missing as it was held at other Trusts. Overall, there were 1,548 data points mandatory for submission based on their CWT pathway type in nine fields of which 95 could not be verified (6%). Of the 1,453 data points that could be verified, 141 had errors in them (10%) (Table 2). 5.2.1 Inaccuracy by cancer pathway Inaccuracies are spread across the Cancer Waiting Times pathways. Treatment related data were the most error-prone with their recording potentially the most misunderstood. Overall, the highest error rate was found in ‘Date of Decision to Treat’ with 27% of cases including an error in this field (57 of 214 cases), ‘Treatment Start date’ included 15% errors (35 of 234) and ‘Treatment Modality’ 5% ( 12 of 239)6 (Figure 5). The proportion of records with data inaccuracies in the remaining six data items are: Date of Diagnosis7 17% (13 of 75 cases); First Seen Date 7% (8 of 108 cases); Referral Received date 8% (9 of 108 cases); Source of Referral 5% (5 of 109 cases); Waiting Time Adjustment (First Seen) 2% (2 of 117 cases); Waiting Time Adjustment (Treatment) 0.4% (1 of 249 cases); 5 While 8% of the patients were treated in other Hospitals, in only 6% of cases was data missing. There is an overlap in these fields and some of the cases with errors in one field e.g. Treatment date may also have errors in another field. 7 Prior to mid-2013 there was no requirement to collect data on Date of Diagnosis, hence there were noticeable inconsistencies this field. 6 20 15th December 2014 Distribution of Errors by Data Items 30% 25% 20% 15% 10% 5% 0% Figure 5: Proportion of Errors by Data Items 5.2.2 Data inaccuracy by tumour site There were data errors across all specialties (Figure 6). In the 139 cases where there was at least one data error, the highest proportions of errors were in Lower GI (24% or 34 cases) and Upper GI (23% or 32 cases), which together account for nearly half of all inaccuracies (47%). Inaccuracies in other specialties included: Urology (17% or 24 cases), Breast (14% or 20 cases), Haematology (8% or 11 cases), Lung (6% or 8 cases), Head & Neck (2% or 3 cases), Skin (2% or 3 cases), Gynaecology (2% or 3 cases), and Cancer of Unknown Primary (< 1% or 1 case). Distribution of Errors by Tumour Site 30% 25% 20% 15% 10% 5% 0% Lower GI Upper GI Urology Breast Others (Haema) Others (Lung) Others (Gynae) Others (H&N) Others (Skin) Others (CUP) Figure 6: Proportion of Errors by Tumour Site 21 15th December 2014 5.2.3 Spread of data errors by tumour site The spread of errors across the pathways for each tumour site is listed in Table 3. It can be seen that six specialties had errors in the majority (>5) of fields relating to the CWT standards. Lower GI had the highest number of inaccurate key data items, with eight out the nine fields audited having errors; Breast and Urology both had inaccuracies in six of the fields; Head & Neck, Haematology, and Upper GI all had errors in five fields; Skin and Lung in three fields), Gynaecology in two fields and Cancer of Unknown Primary in one field. Data Items Where Inaccuracies Were Found by Tumour Site Tumour Site Date of DTT Treat ment Start Date Date of Diagnosis Treatme nt Modality Date Referral Received Source of Referral Date 1st Seen Waiting Time Adj. (1st Seen) Total Yes Waitin g Time Adj. (Treat ment) Yes Lower GI Yes Yes Yes Yes No Yes Yes Breast Yes Yes Yes No Yes Yes Urology Yes Yes Yes Yes Yes No Yes No No 6 No Yes No 6 Head &Neck Yes Yes Yes No Yes Yes No No No 5 Haematology Yes Yes Yes Yes Yes No No No No 5 Upper GI Yes Yes No Yes Yes No Yes No No 5 Lung Yes Yes No Yes No No No No No 3 Dermatology Yes Yes Yes No No No No No No 3 Gynaecology Yes No No No No Yes No No No 2 CUP No No No Yes No No No No No 1 Total 9 8 6 6 5 4 3 2 1 8 Table 3: Distribution of Inaccuracies in the 9 Key Data Items by Tumour Sites Note. Number of cases reviewed - Breast, Lower GI, Upper GI, and Urology (all 50 cases each) and Gynaecology (5), Head & Neck (3), Haematology (16), Lung (16), Cancer of Unknown Primary (CUP) (2), and Skin (8). 22 15th December 2014 5.3 Causes of data errors Data was coded at the end of the tool into categories set by IMT (Table 4). In 40% there is no data error while in 39% data is available but not entered or updated on the CWT. Code Definition No errors The record is found to be accurate and complete in all fields. Data entry errors When CWT and other sources of information, including 40% (100 patient notes, match although the data has not been cases) updated e.g. on another pathway, appointment brought forward, incorrect treatment entry When an incorrect entry appears to have been made as a 11% (28 cases) result of a misunderstanding of the national CWT guidelines. Misinterpretation of the national guidance Operating process issue Potential data manipulation Data with other provider % found (number of cases) 39% (97cases) A substantive error in a process or procedure, as opposed to 3% (7 cases) the recording of that process or procedure. Where there is no valid explanation why CWT has been 2% (4 cases) altered to meet national reporting standards and data does not reflect actual patient experience • There is variance; and/or • There is no valid reason • Other sources give rise for concern e.g. complaints Data available on the CWT but could not be checked against 6% (14 cases) data in the notes which were held elsewhere. Member of staff Clock stop on cancer pathway by any non-practicing clinician working outside the scope of their responsibility 0 (no case) Table 4: Categorisation of data issues To provide a more detailed analysis, where a data inaccuracy was recorded, cases could be ascribed a further level of coding based on cause of the errors. The frequency of these errors is described below and given in Figure 7. 39% (23 of 59 cases) were due to ‘Poor Information Sharing,’ where a discrepancy between CWT and other records arose because information required for accurate record-keeping had not been communicated effectively. This may include a failure to make information available to anyone who might reasonably have use for it or failure to acquire necessary information that is readily available. 39% (23 of 59 cases) were due to ‘Poor document handling,’ where a document had not been processed or used correctly. Instances may include documents mislaid, misread, incorrectly completed, misunderstood, or incorrectly reported. 23 15th December 2014 In 20% (12 of 59 cases) a treatment decision or discussion in the Multidisciplinary Team was mis-recorded as the Decision to Treat Time. National guidance requires the Decision to Treat Time to be the point at which treatment is agreed with the patient. In 2% (1 of 59 cases) there were errors relating to use of Earliest Clinically-Appropriate Date (ECAD) or mis-recording of the DTT, which resulted from confusion over the ECAD for treatment.8 Distribution of Errors (Level Two) 45% 40% 35% 30% 25% 20% 15% 10% 5% 0% Poor information sharing Poor document handling MDT date used in error ECAD Issue Figure 7: Basis of Errors (Level Two) 5.4 Impact of data inaccuracy 5.4.1 Time between recorded and actual dates There were 105 data points on the cancer waiting time pathway (“episodes”) where auditors calculated a difference or variance between the dates in the notes and the CWT tool. These occurred in four fields: referral received date, date first seen, decision to treat, and treatment start date. These data points are listed in Appendix 5. Table 5 gives the size of difference in time between dates recorded in the notes and CWT tool for each field, where a negative number indicates the date in the notes was before that in the CWT tool, and positive number indicates the dates in the notes is after that recorded in the CWT tool. Figure 8 and 9 show the size and distribution of the difference. 8 An ECAD (Earliest Clinically Appropriate Date) can be inserted in the Decision To Treat field within the CWT Tool in accordance with CWT Guidance when a patient is not clinically fit to be on a waiting list i.e. the clock has not started because they are not yet fit to undergo the next activity on their care pathway (para 3.10.15, page 90, CWTs guidance v8.0). The CWT database used the same field to record DTT and ECAD. 24 15th December 2014 Comparing the dates recorded in the notes and on the CWT tool: In most instances where an error was identified the difference between the date recorded on the CWT and the date in the notes was small. In 58% of episodes (61 of 105) the date recorded in the notes was within five days before or after the date recorded CWT Tool. In the majority of episodes (65% or 68 of the 105 data points) the date recorded in the notes was earlier than the date recorded on the CWT Tool. Overall, only in treatment start date was the date recorded in the notes after that recorded in the CWT tool; on average by six days. In the other three fields (referral received date, date first seen, and decision to treat), dates recorded in the notes were, on average, before those recorded in the CWT tool. For 10 patients, dates recorded in the notes were between 10 and 52 days later than those recorded on the CWT Tool, indicating that, for some patients, and especially as the pathways may be cumulative, there would have been unacceptable delays. For all pathways, dates recorded in the notes range from between 77 days before to 52 days after dates recorded in the Cancer Wait Tool. Referral received date 8 episodes of 108 Date first seen Decision to treat Number 9 episodes of 55 episodes of 214 with 108 variance Average -0.9 days - 0.1 day -7.7 days Range -3 +3 -13 +8 -77 +37 of variance (days) Table 5: Summary of Date Variance Identified in 4 Data Fields Treatment start date 33 episodes of 234 6.5 days -14 +52 Total 105 episodes -77 to +52 Difference Between the Dates Recorded in the Notes and CWT Database in 5 Day Intervals Frequency of Occurence 40 35 30 25 20 15 10 5 0 -30+ -26-30 -21-25 -16-20 -11-15 -6-10 -1-5 1-5 6-10 11-15 16-20 21-25 26-30 30+ Number of Days Variance Figure 8: Bar Chart to illustrate frequency of differences between notes and CWT dates in 5 day intervals 25 15th December 2014 Difference in Recorded Date Between CWT and Notes 60 40 20 Days 0 0 20 40 60 80 100 120 -20 -40 -60 -80 -100 Data item or event in cancer pathway Figure 9: Scatter plot of episodes of variance between dates recorded in notes and on CWT tool Table 6 shows variance by specialty and Table 7 by specialty and field. Data is difficult to interpret due to small numbers; however, there was a larger variance in DTT and start date, especially for Lower GI, Lung, Haematology and Urology. Total No. of data items Range of Variance showing variance (days) Average Variance (days) Tumour Site Upper Gastro-Intestinal 24 -25 to 17 -2.8 Lower GI 25 -69 to 52 -1.5 Urology 17 -77 to 44 -4 Breast 17 -28 to 7 -3.9 Haematology 9 -20 to 37 3.2 Lung 7 -7 to 33 3.7 Skin 2 -2 to -2 -2 Head and Neck 3 -17 to 1 -6.6 Gynaecology 1 -18 -18 Cancer of unknown origin 0 TOTAL 105 -77 to 52 -2.1 Table 6: Average variance between the CWT and notes in days across 4 fields by tumour site 26 15th December 2014 Referral Received Date Tumour Site (n = cases of variance) Upper GI (24) Lower GI (24) Urology (17) Breast (17) Haemat (9) Lung (7) H&N (4) Gynae (1) C.U.P Skin(2) Date First seen No of cases Range Aver age No. of cases 2 -2 -2 2 1 -3 -3 5 3 -3 to -2 -2.3 2 -1 to 3 1 1 -1 -1 Range -5 to 7 -13 to 8 Decision To Treat Date Avera ge No. of cases 1 15 -2.2 11 5 5 Averag e No. of cases -5.0 5 -8.7 8 -29.2 9 -6.2 1 3.8 4 -1.3 3 2 4 -25 to 17 -69 to 5 -77 to -6 -28 to 6 -20 to 37 -7 to 4 1 -17 -17 1 -18 -18 5 1 Range 13 4 Treatment Start Date 1 -2 -2 -13 to -77 TOTAL 9 -3 to 3 -0.9 8 8 -0.1 55 to 37 -7.7 Table 7: Variance in days between the CWT and notes by field and speciality 1 33 Range -4 to 17 -5 to 52 -4 to 44 Averag e 2.2 8.9 9.4 7 -14 to 29 -7 to 33 -4 to 1 7 -2 -14 to 52 -2 3.8 10.3 -1.5 6.5 5.4.2 Significance of variance between CWT tool and data in the notes The following statistical analysis considers the significance of any difference in dates between the CWT tool and notes by pathway. If the errors were random it would be expected that the interval to be longer in 50% and shorter in 50%. a) Referral to First appointment (2WW) 91 patients were referred as 2WW. There are 14 patients in whom either the Referral Received Date and/or the Date First Seen was incorrectly recorded in the CWT. Consequently the time to first outpatient attendance would have been incorrectly calculated from the CWT. In four cases the actual time to first outpatient attendance (from notes) was shorter than indicated by the CWT. For these actual performance is better than that indicated by CWT. For 10 patients the actual time to first outpatient attendance (from notes) was longer than indicated by the CWT. For these the actual performance is worse than that indicated by CWT, but in only three patients was there a breach of the 14-day wait that was “missed” by the CWT (in one patient a 16 day wait was recorded as a 14-day wait, in another a 15-day wait was recorded as a seven- day wait and in the third a 17-day wait was recorded as an 11day wait). This suggests a small tendency for errors to favor shorter waiting times on the CWT than recorded in the notes 27 15th December 2014 The proportion of errors that make performance look better than it really is (10/14 (71% (95% CI 42-92%) are not significantly different from 50% indicating this difference is likely to be random. b) Decision to treat to treatment (31-Day) Using the same methodology as above on incorrect dates of Decision to Treat (DTT) and Treatment There were 214 patients with a verified Decision to Treat Date and Treatment Date There were 71 patients (33% (27 – 40%)) where either the DTT or Treatment date was incorrect so the time from DTT to Treatment may be incorrect in the CWT Tool. - In three cases of these, the length of interval from DTT to Treatment is the same overall based on dates recorded in both CWT tool and case notes (i.e. the data errors balance out and do not affect the time from decision to treat and treatment) - In 53 cases (75% (63-84%)), the interval is longer in case notes than recorded in CWT tool, i.e. the patients waited longer in reality than that recorded in CWT tool. - In the remaining 15 cases (21% (12 -32%)), the interval is shorter in case notes than in CWT Tool (i.e. the patients did not wait for as long as wait recorded in the CWT tool) If the errors were random we would expect the interval to be longer in 50% and shorter in 50%, but, as the Confidence Intervals excludes 50%, the data suggest a systematic tendency for errors to favor the system. c) Referral to first treatment (62-Day) Using the same methodology for calculations on incorrect dates of Referral date and First Treatment (1st Treatment) in the 62 day pathway There were 108 patients where the referral and treatment date could be verified There are 24 patients (22% (15-31%) where either the Referral or 1st Treatment date is incorrect. - Of these, in 14 cases 58% ((37 - 78%)), the interval is longer in case notes than recorded in CWT tool (i.e. the patients waited longer in reality than that recorded in CWT tool). - In the remaining 10 cases (42% (22-63%)), the interval is shorter in case notes than in CWT Tool (i.e. the patients did not wait for as long as wait recorded in the CWT tool) Again this suggests a systematic tendency for errors to favor the system, but as the CIs include 50% the tendency is not statistically significant. Overall, there is some evidence of systematic bias making waits seem shorter than in reality in DTT dates recorded. This could have resulted from multiple causes including erroneous interpretation of complex CWT guidance (see sections 5.3 and 7.4), such as a decision recorded as the date the MDT made the recommendation rather than the date the patient made the decision, or even subconscious bias. Data from other Trusts have not been analysed to such a degree as a comparator. The methods of this study did not allow an assessment whether this led to gain, for example, in recording of performance. 28 15th December 2014 5.4.3 Clinical pathways and care Seven cases were referred to the Assurance Panel for potential data manipulation and/or clinical harm. Following in depth discussion and review the following outcomes were agreed (Table 8). Number of cases referred to Assurance Panel 4 cases of “potential data manipulation” 2 Serious Incidents raised by the Retrospective Review Team discussed 1 Case potential clinical harm discussed due to routine referral for red flag symptoms Table 8: Outcome of Review by Assurance Panel Outcome of review No clinical harm and does not meet criteria for referral to police. Agreed variance caused by “potential data manipulation” or “Misinterpretation of the national guidance” Two cases raised for investigation as SIs by the team and supported by Assurance Panel Referred to GP as significant event All four cases identified by the auditors as requiring consideration for “potential data manipulation” were deemed by the Assurance Panel to have involved no clinical harm and therefor did not meet the criteria for referral to police. It was agreed that the data variance was caused by “potential data manipulation or misinterpretation of the national guidance”. Three cases of the 250 patents in the sample were deemed by the Assurance Panel to have potentially experienced clinical harm, of which two had already been referred through the SI process9 by the team. One was a GP significant event10, given that a routine referral was made despite red flag symptoms. No SI was attributed to potential data manipulation. As a result of the review, the Trust is instituting a new policy for repeat Chest X-Rays, where there is lung consolidation and pre-operative anaesthetic assessment, as well as implementing related actions in the Cancer Action Plan. Of the 21 cases treated at other hospitals, in 12 the potential impact on patients’ care could not be established. Cases reviewed by auditors will be shared with other hospitals at the end of the review, unless patient care is affected and more urgency is required. 5.5 Correlation with Patient surveillance log During the CQC investigation, several processes were established to manage patient calls, complaints and significant events. All patients who contacted the Trust in this way were recorded on a “surveillance” log to ensure actions were followed and documented. There were 10 patients on the log also reviewed by the review team (Table 8). The log identified three additional SIs for unrelated issues that would be picked up in other audits e.g. Long Waits and Delayed Diagnoses, bringing the total to five SIs in the study sample. 9 It is important to note that, during the inspection period, the threshold for referral for Serious Incident investigation was lowered. 10 Significant Event Audit - also called Significant Event Review or Analysis - is an increasingly routine part of general practice. It is a technique to reflect on and learn from individual cases to improve quality of care overall i.e. the primary care equivalent to a Serious Incident. http://www.rcgp.org.uk/clinical-andresearch/clinical-resources/quality-improvement/significant-event-audit.aspx 29 15th December 2014 Origin in log Number of cases on the log who called the helpline Number of cases on the log who had made complaints Number of cases on the log who were recorded as SIs Number of cases 6 1 3 cases for unrelated issues that would be picked up in other audits e.g. long delay, delayed diagnosis. Table 9: Correlation between study sample and log of SIs, complaints and calls to the patient helpline related to cancer Of these 10 patients, three each were in Urology and Upper GI, and two each were in Lower GI and Haematology. Of the six patients who had phoned the Trust’s helpline, four required no further action, while the remaining two were referred to the Trust’s Cancer Programme Director. The patient who made a complaint was not deemed to have suffered any potential clinical or emotional harm and, as a result, did not require any further action. 30 15th December 2014 6. Royal Marsden Hospital Review (Appendices 7 and 8) The Royal Marsden Hospital (RMH) was asked to support the Retrospective Review of records for cancer patients at CHUFT by undertaking an independent audit of the review teams’ application of CWT guidance. RMH did this by reviewing a 10% (25) sample of audited patient notes. Of the 25 cases reviewed, RMH agreed completely with Colchester audit results in 18 cases and was unable to agree or disagree with the Trust results in a further two cases because treatment was ultimately carried out at another hospital. Of the remaining five cases, the RMH reviewers disagreed with the Trust’s assessment in at least one CWT data field out of nine per patient (Appendix 1). The purpose of the RMH review was not to identify any potential data manipulation or patient harm but to assess the accuracy of the assessments made by the Trust’s review team. On discussion of these findings, the RMH agreed that: CWT guidance is highly complex and therefore can be open to misinterpretation. Therefore the importance of commissioner-agreed local guidance was emphasised. The differences between the RMH and Colchester assessment were due to technical aspects of CWT guidance and/or the complexity of the pathways reviewed. To maintain high data quality standards providers need to regularly audit their CWT data to identify and rectify any errors. Under these circumstances, the RMH review does not invalidate the findings of the Colchester audit but rather provides a helpful pointer towards the need for greater data accuracy and contextualises the audit, illustrating the complexity of the CWT rules and the need for vigilance in interpreting them appropriately. 31 15th December 2014 7. Areas identified for improvement from qualitative analysis Non-clinical action is defined as action taken to change administrative, technical, or clerical activity over and above the care of an individual patient. Themes from comments relating to cases coded for “nonclinical action” are listed below together with those from the Issues log: 7.1 Poor record-keeping Record-keeping, both electronic and written, is patchy. Records are often incomplete, disordered, illegible, and sometimes inaccurate. The physical state of case notes is often decrepit and binding antiquated. The quality of electronic record-keeping is highly variable across the Trust, reflecting varying IT literacy. 7.2 Decentralised record-keeping Patient records are frequently stored across multiple locations within the Trust, both physically and electronically (e.g. Oncology ‘blue’ notes, radiology notes, separate CNS contact notes). As a result, staff may not know where to find important information, be unable to access it, or may not know that it is held at all. This applies to paper records, but even more so to electronic systems. There are numerous clinical systems in use, apparently with no facility for aggregating their data into a single interface. 7.3 Misunderstanding of Cancer Waits guidance There is evidence of a lack of understanding about Cancer Waits guidance.11 Examples include when pathways start and stop, timing points within pathways, and patients being put on the incorrect pathway including treatment. 7.4 Misunderstanding of “Decision to Treat” In particular, there appears to be greater lack of understanding of what constitutes the Decision to Treat (DTT) date leading to incorrectly recording it as the Outpatient appointment following the correct DTT; the MDT discussion date; the date the patient signed the consent form (rather than the discussion several days previously); the date seen by palliative team (treatment start date); the date of the pre-op assessment; the date of an emergency admission; the date of investigations; and the date of follow-discussion about surgery (rather than when surgery was decided upon). 7.5 Poor information sharing with other providers There have been a significant number of cases in which the full patient record was not available to auditors because information about parts of the pathway was held by another provider or providers. 11 Department of Health, “Cancer Wait Times (CWTs) A Guide” Ver. 8.0 (2011). 32 15th December 2014 8. Discussion and conclusion This audit indicates considerable inaccuracy in the Cancer Waiting Times tool due to multiple reasons across most specialties reviewed. Although the actual error rate of 54% may not be a true picture, as the sample was taken from a selected cohort of specialties with known issues, over half of cases having a data inaccuracy indicate poor data quality and recording. The condition of the health care record is poor. While most errors were relatively minor, and indicate that many patients would have been seen before the date recorded on the CWT, the difference between the dates in the notes and that recorded on the CWT pathway indicate a minority of patients would have experienced unacceptable delays. There is some evidence of systematic bias on statistical analysis of the data which make waits seem shorter than in reality. This could have resulted from multiple causes including erroneous interpretation of complex CWT guidance or even subconscious bias. In the review of individual cases however by the assurance panel, evidence of data manipulation was not identified. Data from other Trusts have not been analysed to such a degree as a comparator. The methods of this study did not allow an assessment whether this led to gain, for example, in recording of performance. The picture is of a problem with data quality. There were pockets of particularly poor data recording including Upper and Lower GI and Urology especially around treatment dates. The size of the variation between CWT and data recorded in the notes is unacceptably large especially in Urology and Lower GI, but also in other specialties. In the sample of 250 cases, there were five that were referred for investigation as Serious Incidents. Two referrals arose from the review and three had been referred previously via the helpline, complaints, or the SI process. These latter three SIs related to aspects of care other than those covered by the Error Rate audit. It should be noted that, during the period of the helpline, the threshold for SI investigations was lowered. As a result of review of these investigations, the Trust has put new policies in place and reinforced aspects of the Cancer Action Plan. As one significant event was as a result of GP referral on to a noncancer pathway, and as 5% of GP cancer referrals were routine, the Trust needs to work closely with primary care to improve early recognition and referral of cancer across the system. The Trust needs to work with other hospitals to improve data sharing as, in 14 of 21 cases of shared care, insufficient data were available to complete the audit. 33 15th December 2014 9. Recommendations The following recommendations have been reviewed by the Cancer Programme Director and associated actions already included in the Cancer Action Plan. 9.1 There should be improved staff training in Cancer Waiting Times (similar to Referral to Treatment (RTT) and greater accountability for record keeping. The Trust needs to recognise the importance of record-keeping and ensure that it is reflected in the job descriptions of those responsible for it. There should be a set of core competencies and an on-going programme of assessment. It is key that staff understand the importance of the data they are recording as part of the Patient’s Clinical Record. Action Taken: As part of the Cancer Action Plan, the Trust is 1 2 3 Developing an e-learning module relating to CWTs guidance which will be used to test understanding of CWT guidance, its application, and competences/skills required. A set of MDT protocols is being developed for the MDT Co-ordinator team which will enable robust cross –cover for MDTCs during periods of absence. As part of the Cancer Action Plan, there is a standing agenda item on the MDT Co-ordinator (MDTC) weekly team meeting for training on the CWTs guidance. The MDTC team determine which aspect of CWTs guidance is covered each week. 9.2 There should be an ongoing process to assess data quality in cancer through a rolling programme of audits reporting to the Board, which also reports on the quality of case notes as well as accuracy of electronic records. 9.3 Any healthcare provider treating a patient should have ready access to that patient’s entire record and history. There should be a robust and clear method for inter-trust communications. This deficiency may be remedied to some extent by the introduction of digital record-keeping across the NHS. Action Taken: As part of the Cancer Action Plan: 1 2 3 The implementation of the Somerset Cancer Registry (SCR) has enabled closer cross-working with neighbouring Trusts. Mid Essex Hospitals NHS Trust has granted CHUFT read-only access to its SCR system, reducing the need for patient data to be transmitted between organisations. Once the roll-out of Phase 2 (MDT Live Data Collection) has been completed, CHUFT will grant Mid Essex Hospitals read only access to its SCR system (estimated for end September ’14). As a minimum, all inter-trust referrals should be accompanied by the Department of Health (DH) defined Minimum Data Set for Cancer. An SCR module is being developed, which will generate inter-trust referrals with the Minimum Data Set. CHUFT is working with the SCN (Strategic Clinical Network)/ ECRIC (National Cancer Registration Service, Eastern Office) to agree an Essex-wide inter-trust referral policy which sets out a clear and robust methodology for transmission of data between Trusts. This policy is based on the Anglia Clinical Network Intra Trust Referral Policy, which has been adopted by Ipswich Hospital. 9.4 Patient paper records should be stored together wherever possible. There needs to be a central inventory of clinical systems with a view to their integration wherever possible. Departments and 34 15th December 2014 service areas need to have a clear map of where they store data and who has access rights. The patient’s entire record should be accessible via a single interface. The Trust should consider implementing an electronic patient record as a high priority. 9.5 This report should be fed back to clinical teams to consider the impact of data quality and develop an action plan. Check reports have already been implemented between the MDT and Business Informatics teams – to flag differences between PAS and the SCR. 9.6 Breaches should be continually monitored, especially around treatment dates, and daily reporting put in place to escalate patients waiting, as well as feedback on data quality to the Divisions and specialties. 9.7 The NAO recommends that Clinical commissioning groups and trusts should work together to impress on patients their rights and responsibilities. Almost all patients in the small sample we interviewed were unaware of the 18-week maximum waiting time and the implications if they failed to attend their appointment – although patients failing to turn up for appointments is a longstanding challenge which should not be underestimated.12 The Trust and CCG should work with patients to understand their rights and responsibilities around cancer waiting times. 12 Department of Health, “NHS waiting times for elective care in England,” HC964 Session 2013-14 23 January 2014, National Audit Office, p. 11. 35 15th December 2014 Appendix 1: Study Protocol Study Title: To determine the error rate of the Cancer Waiting Tool Chief Investigator: Dr. Christine Macleod Medical Director Essex Area Team NHS England [email protected] Sponsor: Incident Management Team, Essex Area Team, NHS England Confidentiality Statement This document contains confidential information that must not be disclosed to anyone other than the Sponsor, the Investigator Team, host organisation, unless authorised to do so. Background to the project The Care Quality Commission (CQC) report into Cancer standards at Colchester Hospital University Foundation Trust (CHUFT), published on 5 November 2013, identified a number of failings in Cancer Services in the Trust including “in 22 cases people were at risk of or did experience delays in their care which could have a negative impact on the person using the service.” An analysis by the NHS IMAS Intensive Support Team of cancer data indicated that 40,489 patients had undertaken 66,997 cancer pathways at CHUFT from 1 April 2010 to date. The Cancer Waiting Tool (CWT) data do not all fully reconcile with data in other systems such as PAS. There may be valid reasons for this variance which can be clarified by this audit. CHUFT cancer waiting time performance reporting is based on output from the CWT. If the CWT data are inaccurate, the actual performance of CHUFT against the National Standards for Cancer Waiting Times may also be inaccurate. Aim of the study The aim of this study is to assess the historical and current accuracy of the data held on the Cancer Waiting Tool so that accurate Trust cancer waiting time performance can be assured. Objectives Objectives Outcome Measures/Endpoints Primary Objective 1 To determine the historical and current accuracy of the data on the Cancer Waiting Tool % data accuracy of CWT April 2010 – November 2013 % data accuracy of CWT now in 2014 Secondary Objectives 2 To estimate the impact of the CWT data accuracy on cancer waiting time performance reporting Estimated impact on historic / current performance reporting 36 15th December 2014 Experimental design and methods (including statistical analysis) 1. Randomised sample of 250 case notes, for patients on the 31 or 62 day pathway between 1 April 2010 and 31 December 2013 (as advised by Dr. Paul Pharoah, Public Health & Primary Care, University of Cambridge) 2. Sample stratified by the four cancer pathways identified by the CQC inspection, namely Urology, Breast, Upper GI, Lower GI - 50 case notes from each pathway, 3. In addition, a random sample of 50 case notes across the other cancer pathways, 4. Scope for additional second stage sampling of specific cancer pathways if serious problems identified in first stage, 5. Identification of key metrics from case notes: • • • • • • • • • referral source (GP, consultant, screening, other) referral date date referral received (same as date referral received for 62-day pathway and date of decision to treat for nonurgent pathways) date first seen (where patient was a GP target referral or a screening referral) diagnosis date (if available retrospectively, as not ‘mandatory’ until 2013) decision to treat date date of First Definitive Treatment (FDT) as per national guidance details of treatment / assessment of appropriateness of whether recorded treatment was a recognised First Definitive Treatment (in line with national guidance – may require independent specialist input to assess specialist clinical management) date of subsequent treatment(s) or other appropriate actions e.g. pauses and recording of DNA 6. Compare case note metrics: • • • • with data on CWT with data on the CNS database(s) with data on Open Exeter against National Standards for Cancer Waiting Times 7. Analyse error rate(s) 8. Further sampling of case notes as indicated 9. Repeat audit of a sample of notes to be undertaken in late February / March 2014 to assess any difference in the error rate post the Trust interventions in late 2013 (precise methodology to be confirmed following results of initial audits as above) Ethical considerations Ethical considerations and information governance issues that need attention include: data confidentiality 37 15th December 2014 The essential elements of this review are technically audits and can be undertaken by those who have a legitimate clinical relationship with the patient. However given the amount of scrutiny that the Trust is under, the protocols for such reviews will be agreed with NHS England, who will also provide external assurance, validity and scrutiny of the process (including the checking of the accuracy of any reviews under the section 251 application). Resources and costs 1. Decide on funders. 2. Scale of work - randomisation and pulling of case notes = 10 minutes per case note = 42 hours / 5 days 3. Scale of work - 40 minutes per case note = 170 hours / 21 days = £10k @ 500 per day, if at clinical consultant rate. 4. Scale of work - analysis of data = estimate 3-4 days = £1-2k 5. Scale of work - report = estimate 1 day = £0.5k 6. Cost - approximately £15k but need clarification of costs for point 2 above. 7. Cost of 2014 re-audit sample to be determined 38 15th December 2014 Appendix 2: Data on Cancer Survival, Mortality and Referrals Table 1: NE Essex (NEE) CCG one-year survival index (%) within normal range for all cancers combined for adults (aged 15-99 years) Source: ONS and National Cancer Registry 75.0 73.0 71.0 69.0 67.0 England 65.0 63.0 Essex 61.0 North East Essex 59.0 57.0 55.0 Source: ONS and National Cancer Registry Table 2: NEE CCG one-year survival index (%) for all cancers combined compared to Essex CCGs Source: ONS and National Cancer Registry Table 3: NE Essex 1 and 5 year survival rates for cancers within key specialties are all within the normal range 1 year survival 5 Year survival Breast Normal range Normal range Colorectal Normal range Normal range Lung Normal range Normal range Prostate Normal range Normal range Urological Normal range bladder, kidney, Normal range bladder, kidney, testicular testicular 39 15th December 2014 Source: PH England data and knowledge gateway: National Cancer Intelligence Network e-atlas Table 4: NE Essex mortality 2008-2012 by cancer type all within normal range except prostate, lung and stomach which are significantly lower than UK average Mortality for 2008-2011 and 2009-2012: National Cancer Intelligence Network Normal range Normal range Normal range Persons significantly low, female low, male normal limits Significantly low Normal range bladder (male & female), kidney (male & female), testicular Normal range cervical, uterine, vulval, ovarian cancer (male & female) Normal range non-Hodgkin’s lymphoma (male & female) Normal range oesophagus (male & female) All cancers Breast Colorectal Lung Prostate Urological Gynaecological Haematological Upper GI Stomach (all person low, male & female normal) Skin Normal range melanoma (male and female) Brain Normal range (male and female) Melanoma Normal range (male and female) Pancreas Normal range (male and female) PH England data and knowledge gateway: National Cancer Intelligence Network e atlas NE Essex Table 5: Total number of patients referred for 2WW is increasing although not translating into increased numbers diagnosed and treated Not Treated Treated 1800 1600 1400 1200 1000 800 600 400 200 01/08/2013 01/06/2013 01/04/2013 01/02/2013 01/12/2012 01/10/2012 01/08/2012 01/06/2012 01/04/2012 01/02/2012 01/12/2011 01/10/2011 01/08/2011 01/06/2011 01/04/2011 01/02/2011 01/12/2010 01/10/2010 01/08/2010 01/06/2010 01/04/2010 01/02/2010 01/12/2009 01/10/2009 01/08/2009 01/06/2009 01/04/2009 0 40 15th December 2014 Appendix 3: Methodology for Obtaining Sample Data 1. Data Source: Historic record of Open Exeter quarterly submissions 2. Date Parameters: Data submitted to Open Exeter from Quarter One 2010/11 to Quarter Three 3 2013/14 (i.e. April 2010 to December 2013) 3. Summary of procedures: i. All the records from the specified period (i.e. data submitted to Open Exeter) were extracted and copied in to an Excel document. ii. The data were then filtered to the specific tumour sites required for the audit and copied to a selection work sheet. iii. The selection worksheet contains formulas required to randomise the sample and to identify any patients selected for previous audits. iv. A randomised number was generated against each record and records selected for previous 2013 audits flagged by their NHS numbers or PAS numbers for exclusion. v. The randomised sample was then sorted and the first 50 for each of the 5 cohorts were selected as the sample data. vi. An additional 16 patients were selected to bring the sample size to 266 records as a contingency in case notes could not be found on time. vii. Following an initial assessment by the RRT, the 266 data sample was stratified down in order to limit the sample size to 250 records in line with the audit’s requirement. This was done by removing a total of 27 records using the following criteria: (a) 14 excess records yet to be reviewed in two cohorts (i.e. eight in Breast and six in Lower GI); (b) seven duplicate records; (c) two records where the pathway type could not be determined; (d) four records where pathways commenced outside the specified date parameters. This resulted in a total of 239 unique records leaving a shortfall of 11 records. viii. To augment for observed shortfall, 11 additional records were selected at random using the same methodology described in steps ii to vi above in the 3 cohorts where the number of records was less than 50, viz Upper GI- 4, Urology- 3, other tumour sites-4. ix. At some point, the RRT ran out of notes to audit and, to avoid any delay, an additional 53 records were drawn at random, proportionate to the outstanding number of records to be audited in each of the five cohorts. x. Once the threshold of 50 was reached in a cohort, case notes in that cohort were no longer audited. Notes on analysis 1. Patient age was determined by how old they were on the day they commenced treatment rather than their age on the day their referral was received. This is because the date of receipt of referral was not recorded for many of the patients in the sample. 2. Based on Cancer Outcome and Service Dataset (COSD) and CWT reporting guidelines, age distribution of patients on cancer pathways is usually grouped into “Children” (aged under 16), “Teenagers and Young Adults” (TYA – aged 17-24) and “Adults” (aged 25 and above). To enable a robust analysis however, the Office of National Statistics (ONS)’s 5-year age distribution model was adopted. 3. None of the patients in the sample data was aged under 25 and so the cancer pathways in young adults were not reviewed. 4. Certain elements of the shared pathway could not be verified for accuracy since the information is held in other Trusts. 41 15th December 2014 5. In line with national statutory CWT reporting standards, certain data items on patients on routine and subsequent treatment pathways are not mandatory for national submissions; as a result many of these were not reviewed in the audit. 42 15th December 2014 Appendix 4: Audit Tool and Fields Field Pt Identifier Pt in other Audits? cwtMainId Hospital Number First Name Second Name NHS Number Patient DoB Pt Age Today Pt Age @ Tx Patient Sex Death Date Cancer Type Description of source of referral Priority Type Description Cancer Referral to Treatment Period Start Date Start Variance (Days) Cancer Type 2 Week Wait Date first seen Date first seen variance (days) Cancer treatment period start date Decision To Treat Variance (Days) 1st or 2nd definitive treatment Hospital of Treatment Treatment start date Treatment start date variance (days) Treatment modality Existing Complaint? Existing Serious Incident? Contacted Helpline? Description of source of referral Cancer Type (Re-grouped) In SI list? Pt Age Group Cancer Type (Re-grouped) Category Auto-populated Auto-populated Auto-populated Auto-populated Auto-populated Auto-populated Auto-populated Auto-populated Auto-populated Auto-populated Auto-populated Auto-populated Auto-populated Auto-populated Auto-populated Auto-populated Auto-populated Auto-populated Auto-populated Auto-populated Auto-populated Auto-populated Auto-populated Auto-populated Auto-populated Auto-populated Auto-populated Auto-populated Auto-populated Auto-populated Auto-populated Added later for analysis Added later for analysis Added later for analysis Added later for analysis Revised Pathway based on treatment Event Patient on Correct Pathway/ Cancer type? Patient on Correct Pathway?- for analysis Patient on Correct Pathway (Comments) Referral Source Correct? (Code One) Referral Source Correct? (Code One)- for analysis Added later for analysis Auditor-populated Added later for analysis Auditor-populated Auditor-populated Added later for analysis 43 15th December 2014 Field If no, enter correct source If no, comment (Referral Source) Date Receipt of Referral Correct? (Code One) Date Receipt of Referral Correct? (Code One)- for analysis Date Receipt of Referral if not correct Referral Receipt Date Variance (Code Two) Referral Receipt Date Variance (Code Three) Referral Receipt Date Variance (comment) First seen date correct? (Code One) First seen date correct? (Code One)- for analysis If no, actual First Seen Date Reason for date first seen variance (Code two) Reason for date first seen variance (Code three) Reason for Date First Seen Variance (comment) Date First Seen Adjustment made? Legitimate Date First Seen adjustment (Code one) Legitimate Date First Seen adjustment (Code one)- for analysis Legitimate Date First Seen adjustment (Code two) Legitimate Date First Seen adjustment (Code three) Number of days' Date First Seen adjustment Legitimate Date First Seen adjustment ('No') - comment Date of diagnosis (NB may not be available pre-2013) Date of diagnosis correct? (Code One) Date of Dx correct? (Code One) - for analysis Reason for Diagnosis Date variance (Code two) Reason for Diagnosis Date variance (Code three) Reason for Diagnosis Date Variance - Comment Decision to treat date correct? (Code One) If no, correct Decision To Treat date Decision to treat date correct (Code two) Decision to treat date correct (Code three) Decision to treat date correct (Comment) Treatment start date correct? (Code One) Actual treatment start date Treatment start date correct? (Code two) Treatment start date correct? (Code three) Treatment start date correct? (Comments) Treatment start date adjustment made? Legitimate treatment start date adjustment (Code one) Legitimate treatment start date adjustment (Code two) Legitimate treatment start date adjustment (Code three) Number of days treatment start date adjustment Legitimate treatment start date adjustment ('No') - comment Category Auditor-populated Auditor-populated Auditor-populated Auditor-populated Auditor-populated Auditor-populated Auditor-populated Auditor-populated Auditor-populated Auditor-populated Auditor-populated Auditor-populated Auditor-populated Auditor-populated Auditor-populated Auditor-populated Added later for analysis Auditor-populated Auditor-populated Auditor-populated Auditor-populated Auditor-populated Auditor-populated Auditor-populated Auditor-populated Auditor-populated Auditor-populated Auditor-populated Auditor-populated Auditor-populated Auditor-populated Auditor-populated Auditor-populated Auditor-populated Auditor-populated Auditor-populated Auditor-populated Auditor-populated Auditor-populated Auditor-populated Auditor-populated Auditor-populated Auditor-populated 44 15th December 2014 Field Recorded treatment modality correct? (Code One) If no, state recorded treatment Planned 2nd/subsequent treatment? (Code One) Planned 2nd/subsequent treatment? (Code One) -for analysis 2nd/subsequent treatment received? Auditor comments on pathway Who else was involved in audit? Final code Possible Impact (Code) General comments Comment Summary DT Code 1 Code 2 Action(s) agreed (Code) Assurance Panel Name of Auditor Record Locked? Date completed Actions taken By Whom Date Supplementary Evidence Required Category Auditor-populated Auditor-populated Auditor-populated Added later for analysis Auditor-populated Auditor-populated Auditor-populated Auditor-populated Auditor-populated Auditor-populated Added later for analysis Added later for analysis Added later for analysis Auditor-populated Auditor-populated Auditor-populated Auditor-populated Auditor-populated Added later for analysis Added later for analysis Added later for analysis Added later for analysis Added later for analysis Added later for analysis Added later for analysis Added later for analysis Added later for analysis Added later for analysis Revised Pathway based on treatment Event Priority of Ref Added later for analysis Added later for analysis 2 Week Wait LENGTH: Ref-1st Seen Added later for analysis 31Day LENGTH: Decision To Treat-treatment 62Day LENGTH: Ref-1st treatment Breach? W/E Date Completed ?variance (all) ?variance (to include) Added later for analysis Added later for analysis Added later for analysis Added later for analysis Added later for analysis Added later for analysis Variance in Date? Date of Receipt of Referral Variance Added later for analysis Added later for analysis Date of clinician review Clinician Comments Clinician Name Previous Action Code Previous Final Code Previous Harm Code 45 15th December 2014 Field Date 1st Seen Variance DDT Variance D treatment Variance Category Added later for analysis Added later for analysis Added later for analysis Variance 2 Week Wait LENGTH: Ref-1st Seen 31D 62D Ayo's Comments Added later for analysis Added later for analysis Added later for analysis Added later for analysis Impact on accuracy of CWT Performance Audit Status Refer McStay Hospital of Treatment Combined target breached Pt Age Added later for analysis Added later for analysis Added later for analysis Added later for analysis Added later for analysis Added later for analysis 46 15th December 2014 Appendix 5: Glossary Level One Codes: 0 Correct: The CWT entry is verifiable from case notes or other internal (Trust) sources. This requires no further specification 1 Incorrect: The CWT entry conflicts with the information contained in case notes or other internal sources. This can be further specified with seven Level Two codes: A Entry not in accordance with CWT Guidance B Entry not in accordance with Trust policy C Typographical error D Process error E Unexplained F Other G Suspected manipulation. 2 Not on CWT: A given entry is missing from the CWT database. There is no further specification. 4 Unable to Verify CWT: While it cannot be demonstrated to be incorrect, the CWT entry cannot be corroborated from case notes or other internal sources. This can be further specified with two Level Two codes, ‘H Data Missing’ and ‘I Data with other provider’ (See below). Level Two Codes: The following codes (A-G) are used to further specify the Level One code ‘1 Incorrect’. A Entry not in accordance with CWT guidance: An incorrect entry appears to have been made as a result of a misunderstanding of the national CWT guidelines. For Decision to Treat only, this can be further specified with three Level Three codes: iii MDT date used in error iv ECAD issue vi Nullified by concurrent treatment B Entry not in accordance with Trust policy: An incorrect entry appears to have been made as a result of a misunderstanding of Trust policy. There is no further specification. C Typographical error: An error that appears to result from carelessness, failure of keyboard skills, oversight, etc. It must be apparently unintentional and not motivated by a misunderstanding of guidance or policy. There is no further specification. D Process Error: A substantive error in a process or procedure, as opposed to the recording of that process or procedure. This can be further specified with two Level Three Codes ( I Poor Information Sharing and II Poor Document Handling). 47 15th December 2014 E Suspected manipulation: A discrepancy appears to result from a deliberate attempt to misrepresent the patient pathway as recorded in case notes or other non-CWT sources. Although it will most likely be an alteration to the CWT that improves apparent performance, it does not necessarily have to be. There is no further specification. F Unexplained: An error that cannot be otherwise explained. There is no further specification. G Other: This code is used for anything that does not fit under any of the Level Two codes. There is no further specification. The following two Level Two codes are used to further specify the Level One code ‘3 Unable to verify CWT’. H Data missing: Corroborating information, which should be available in Trust-held patient records, cannot be found. There is no further specification. I Data with other provider: Corroborating data can reasonably be expected to be held by another provider and so is unavailable to the auditor. There is no further specification. Level Three Level Three codes further specify Level Two codes. The first two codes are available to all variance fields. The remaining three apply to the Decision to Treat Time (DTT) variance field, only. i) Poor information sharing: A discrepancy between CWT and other records arises because information required for accurate record-keeping has not been communicated effectively. This may include a failure to make information available to anyone who might reasonably have use for it or failure to acquire necessary information that is readily available. ii) Poor document handling: A document has not been processed or used correctly. Instances may include documents mislaid, misread, incorrectly completed, misunderstood, or incorrectly reported. Decision to Treat Time-only codes iii) MDT date used in error: A treatment decision or discussion in the Multidisciplinary Team is mis-recorded as the Decision to Treat Time. National guidance requires the Decision to Treat Time to be the point at which treatment is agreed with the patient. iv) ECAD issue: Any mis-recording of the DTT that results from confusion over the Earliest Clinically-Appropriate Date for treatment. vi) Nullified by concurrent treatment: 48 15th December 2014 The Date First Seen (DFS) coincides with the date of first treatment; e.g. the excision of a polyp or a diathermy carried out at the same time as a cystoscopy. Final codes The RRT apply the following ‘Final Codes’ to each audited record as a whole. Where several failings are present in a record (i.e. several fields with variance are found), the auditors will apply the code corresponding to the worst failing found. This assumes that the codes ascend in degree of gravity (i.e. a misinterpretation of guidance is more serious than a data entry error and potential data manipulation is worse than a member of staff working outside their responsibility). So, for example, if a record contains instances of typographical errors and operating process errors, code 3 would be applied to the record as a whole due to process errors being more serious than typographical mistakes. 0. No data inaccuracy: The record is found to be accurate and complete in all fields. 1. Data entry error: When CWT and other sources of information, including patient notes match although the data has not been updated e.g. on another pathway, appointment brought forward, incorrect treatment entry. 2. Misinterpretation of national guidance: For example, pausing the pathway, when the pathways have been changed or recorded that is consistent with national guidelines. 3. Operating process issue: A substantive error in a process or procedure, as opposed to the recording of that process or procedure. 4. Member of staff working outside the scope of their responsibility: Clock stop on cancer pathway by any non-practicing clinician. 5. Potential data manipulation: Where there is no valid explanation why CWT has been altered to meet national reporting standards and data does not reflect actual patient experience • There is variance; and/or • There is no valid reason • Other sources give rise for concern e.g. complaints 6. Data with other provider: Data unavailable as a result of treatment occurring at other hospital. 49 15th December 2014 Appendix 6: Variance (in days) for Data Items with Incorrect Dates Variance (No. of Days) -77 -69 -42 -28 -25 -23 -20 -18 -17 -14 -13 -12 -10 -8 -7 -6 -5 -4 -3 -2 -1 1 2 3 4 5 6 7 8 14 15 17 29 33 37 43 44 52 Grand Total Date First Seen Decision To Treat Date 1 1 Referral Received Date Treatment Start Date 1 1 1 1 1 1 1 1 2 1 2 3 3 2 2 3 1 1 1 1 1 3 2 10 5 2 1 1 1 3 3 2 3 1 1 1 1 2 6 5 3 1 2 1 1 1 1 1 1 1 1 1 1 8 55 9 1 1 1 33 Grand Total Percentage 1 1 1 1 1 1 1 1 2 2 3 3 3 3 3 3 2 4 6 6 20 10 5 3 1 4 1 2 1 1 1 2 1 1 1 1 1 1 105 1% 1% 1% 1% 1% 1% 1% 1% 2% 2% 3% 3% 3% 3% 3% 3% 2% 4% 6% 6% 19% 10% 5% 3% 1% 4% 1% 2% 1% 1% 1% 2% 1% 1% 1% 1% 1% 1% 50 15th December 2014 Appendix 7: The Royal Marsden NHS Foundation Trust independent review of Colchester University Hospital NHS Foundation Trust Retrospective Review Team audit results Introduction The Royal Marsden (RMH) was requested by Colchester University Hospitals NHS Foundation Trust (CHUFT) to support their Retrospective Review of records for cancer patients at CHUFT following the CQC review by undertaking an independent audit of the application of the CWT guidance by the Retrospective Review Team (RRT). Methodology A sample of 25 patient notes were provided by the RRT for review by RMH team RMH used the same principle as external auditors (Deloitte) use in mandatory audits of Cancer Waiting Time data in all NHS Trusts’ Quality Accounts. RMH undertook this audit of records at CHUFT on 24th April and 1st May 2014. RMH carried out a three phase process: o First review was carried out ‘blind’ (i.e. with no reference to the Retrospective Review Team (RRT) audit results). This process reviewed electronic and paper records provided by CHUFT and included: Hospital case notes , including nursing notes, clinic letters, oncology ‘blues’, and radiotherapy data Patient Administration System o Second review compared RMH audit results to the RRT audit results o The RMH auditor presented her findings to an RMH validation team consisting of the Director of Performance & Strategy Implementation, the Service Manager for Performance and the Head of Information Findings RMH has set out its findings case by case in the attached, which outlines whether RMH agrees with the RRT audit results, the rationale for RMH’s findings and reference to the national guidance. Of the 25 cases RMH agreed with the RRT audit results in 18 cases. RMH disagreed in five cases and in two cases RMH was unable to audit due to the treatment having been carried out at another hospital. In the five5 cases where RMH disagreed with the RRT audit: o two cases should not have had a new CWT record – in one case this is due to specific guidance for that other tumour type, in the other the patient was continuing on active surveillance o In the other three cases RMH disagreed with at one or more of the Trust’s CWT data fields Nicky Browne Director of Performance and Strategy Implementation The Royal Marsden NHS Foundation Trust 25 June 2014 51 15th December 2014 Appendix 8: Marsden review of cases for Colchester Hospital NHS Foundation Trust as part of retrospective review – cases where there was disagreement with the Trust Patient CWT data identifier fields 1 Agree With Retrospective Review (Y/N) Rationale for decision Guidance reference Agree N Referral received date Date first seen Treatment type Referring tumour type ENT was the referring condition, haem was the final diagnosis but haem was recorded as the referring condition. Biopsy effectively removed the tumour so counts as FDT as per GFOCW v8.0 section 3.9.24 Disagree Referring tumour type Clock stop. Clock stop - TWR clock start should have been 13/3/12 not 14/3/12. Correct treatment reported but marked as an error by the RRT. Patient had excision biopsy, intention was to remove tumour. Clearly documented that tumour was effectively removed by the excision biopsy. Local guidance needed DTT 2 Agree N Referral received date Referring tumour type Date first seen Clock stop Treatment type Disagree DTT DTT - as diagnosis was unknown at time of procedure, local guidance should be agreed with commissioners regarding where the DTT should be taken as national guidance does not specify. DTT - Both our review and RRT review indicate the OE DTT to be incorrect. However we disagree on the date (does not affect breach status of patient). RRT date 16.4.12 RMH audit 13.4.12 "Partial excision/debulking of a tumour (but not a biopsy for diagnostic or staging purposes unless it effectively removed the tumour even if margins are not clear". Treatment as per GFOCW v8.0 section 3.9.24 52 15th December 2014 3 4 5 Agree treatment type Disagree DTT clock stop Agree DTT treatment type Disagree clock stop N Review team DTT 16/3/12 clock stop 18/3/12 RMH DTT 13/3/12 clock stop 16/3/12. Treatment as per GFOCW v8.0 section 3.9.24 N Treatment as per GFOCW v8.0 section 3.9.24 Disagree Entire treatment N Additional information required to clarify the clock stop. Based on the information available we disagree with the RRT. RT start date not recorded. Treatment start date ie clock stop ascertained from RT end date and working back by number of fractions Clock stop for Active Surveillance should not have been reported as patient was already on Active Surveillance. Patient had been on AS since 2011 no need to record in 2013 Treatment as per GFOCW v8.0 section 3.9.24 53 15th December 2014